Lundbeck had a patent for citalopram for the treatment of depression, which it marketed in Australia under the name Cipramil
Citalopram is a chiral molecule: it can exist in two isomeric forms; a (+)-enantiomer and a (-)-enantiomer. The two forms have the same chemical structure, but they are mirror images. At its priority date, the relevant skilled addressees would have understood that the compound was a racemate or racemic mix consisting of both the (+)-enantiomer and the (-)-enantiomer.
Subsequently, Lundbeck discovered a way to make the (+)-enantiomer in isolated or pure form and, even better, it was this enantiomer that contributed the therapeutic effect of citalopram. It obtained a further patent, claim 1 of which was for:
1. (+)-1-(3-dimethylaminopropyl)-1-(4’-fluorophenyl)-1,3-dihydroisobenzofuran-5-carbonitrile and non-toxic acid addition salts thereof.
The Full Federal Court by 2-1 (Bennett and Middleton JJ; Emmett J dissenting) has upheld the validity of this claim in the face of the prior patent for citalopram.
If resort could be made to the body of the specification, it was clear that the skilled addressees would have understood that claim was a claim to the enantiomer in its pure, isolated form and not to the product as part of a racemic mix.
Emmett J considered the claim was clear and unambiguous. Accordingly, there was no warrant to resort to the body of the specification. Bennett and Middleton JJ, on the other hand, accepted that the trial judge was entitled to accept evidence about how the skilled addressee would have read the claim. Bennett J approved the approach taken by Dr Barker, as delegate for the Commissioner, in Emory University v Biochem Pharma:
 …. He drew a distinction between a reading of the claim in the context of the specification to understand what the claim is talking about (as did Burchett J in International Business Machines Corporation v Commissioner of Patents  FCA 625; (1991) 33 FCR 218), where the whole thrust of the specification makes a limitation clear, and impermissibly importing a limitation from “mere comments” in the description. ….
Her Honour further explained that the citalopram patent didn’t anticipate the claim:
193 The prior citalopram patent described the racemate. It did not describe the pure or isolated (+)-enantiomer. There is no anticipation unless the disclosure of the racemate was, to the skilled addressee, a disclosure of the (+)-enantiomer. As the primary judge pointed out at , the skilled but non-inventive addressee would have understood that (+/-)-citalopram consisted of the (+)-enantiomer and the (–)-enantiomer and would have been able to identify the formulae for the S and R enantiomers but would not have known in the absence of experimentation which was the (+)-enantiomer and which the (–)-enantiomer. As his Honour said, these facts would not point specifically to the independent existence of the enantiomers. They did not disclose an invention which, if performed, would necessarily infringe the Patent.
194 It is the case that the skilled addressee knew that the racemate could be resolved into the enantiomers but there was nothing to tell him or her to do so. Further, the prior citalopram patent was silent as to the means of obtaining the enantiomers and there were different methods available to try to do so. There were no clear and unmistakable directions to obtain the enantiomers. Some of the available methodology may have been successful, other methods may not.
That seems rather to qualify the force of the earlier point. One might even think, with respect, that it limits the patentable subject matter to some particular method of making the purified, isolated (+)-enantiomer, rather than the substance per se.
In contrast to this approach, however, Lundbeck lost its attempt to get the term of the second, escitalopram patent extended pursuant to s 70 of the Patents Act.
The problem for Lundbeck here was that s 70(2) referred to a patent for a pharmaceutical substance per se, but s 70(3) and (5) and s 71 did not. The Full Court found that s 70(2) operated to identify the subject matter of the extension application – the substance per se here being escitalopram.
Because the later sub-sections did not refer to the substance per se, when it came to working out the timing for making the application for an extension of term, the question was (in this case) the date when goods containing the substance (not the substance per se) were first included in the Australian Register of Therapeutic Goods.
Citalopram, of course, contained the (+)-enantiomer and it had been registered in the ARTG. It was first included in the ARTG on 29 December 1997. The application to extend the term of the escitalopram patent was made almost 6 years later – on 22 December 2003.
Section 71(2), however, required the extension application to be made within 6 months of the patent being granted or “the date of commencement of the first inclusion in the [ARTG] of goods that contain … the pharmaceutical substance referred to in s 70(3)” (i.e., within 6 months of the inclusion in the ARTG of citalopram).
The majority also found that claim 5 was invalid for insufficiency. However, Lundbeck did succeed insofar as Alphapharm was found to have infringed claims 1 and 3 of the escitalopram patent, presumably up to the date of expiry of the patent on 13 June 2009.